How to set up WW3 within UFS limited area domain configurations?

[XiaSun-Atmos]

Hi,
What files are needed for running UFS regional domain (e.g. RRFS) with coupled WW3?
Thanks!
Xia Sun-CIRES/GSL

[JessicaMeixner-NOAA]

Hi @XiaSun-Atmos that's a great question! We'd first need a grid that covers the region of interest that matches RRFS. Otherwise the set-up would be similar to how WW3 is set up for other regional domains such as HAFS (there are UFS regression tests for that we could follow). What region of RRFS domain are you looking to cover with a wave grid?

[XiaSun-Atmos]

Thank you @JessicaMeixner-NOAA for your quick response! The ultimate domain would be on the east coast for WFIP3 using RRFS. Below is the domain info we made for the UFS LAMs for use in ufs srw app.

Center Lon: -70.8
Center Lat: 40.6
Horizontal dimension X: 700
Horizontal dimension y: 600
Grid distance: 1km
I checked the hafs regression test and changed the suite to FV3_HRRR, suitable namelist, and corresponding diag/field tables for RRFS. It gave errors like missing connections of variables listed in the nems_configure file.

[JessicaMeixner-NOAA]

@XiaSun-Atmos there are several that use regional domains, maybe try tp2.4?

[XiaSun-Atmos]

Thanks @JessicaMeixner-NOAA . tp2.4 provides good examples for my test. I started a test for rrfs domain. The mesh_rrfs.nc file is made using the scrip.nc file generated from ww3_grid and command of ESMF_Scrip2Unstruct. I feed ww3_grid with lat.dat and lon.dat from rrfs with 3-km domain.
When running the model, the err msg is:

511:  *** WAVEWATCH III ERROR IN W3FLDO :
511:      INCOMPATIBLE GRID DATA :         901        441          1
511:                   SHOULD BE :        1799       1059          2

Does the grid data here mean wind.ww3 (wind.natl_6m) which I copied from @[INPUTDATA_ROOT_WW3]/wind.natl_6m ./wind.ww3?
Should I prepare wind.ww3 using mod_def.ww3 corresponding to rrfs domain? Thanks!

[JessicaMeixner-NOAA]

The wind.natl_6m is specific to the corresponding 'natl_6m' grid. You will have to regenerate wind.ww3 data for your coupled model if you'd like to use wind input from another source that's not covered by your atm model domain in the coupled model or if you're running WW3 standalone.

[XiaSun-Atmos]

@JessicaMeixner-NOAA , I tried to make the wind.ww3 file for my domain. The model crashes with this err:

  0:   ... DONE microphysical lookup tables
184: [h4c35:215431:0:215431] Caught signal 8 (Floating point exception: integer divide by zero)

I ran WW3 offline, it runs and generates outputs.
My ww3_grid.nml is attached. Did you spot anything not right? Also, how should I define the &INBND_POINT_NML?
Thank you so much!

! -------------------------------------------------------------------- !
! WAVEWATCH III - ww3_grid.nml - Grid pre-processing                   !
! -------------------------------------------------------------------- !

! -------------------------------------------------------------------- !
! Define the spectrum parameterization via SPECTRUM_NML namelist
! -------------------------------------------------------------------- !
&SPECTRUM_NML
  SPECTRUM%XFR       =  1.1
  SPECTRUM%FREQ1     =  0.0373
  SPECTRUM%NK        =  32
  SPECTRUM%NTH       =  24
/

! -------------------------------------------------------------------- !
! Define the run parameterization via RUN_NML namelist
! -------------------------------------------------------------------- !
&RUN_NML
  RUN%FLCX         =  T
  RUN%FLCY         =  T
  RUN%FLCTH        =  T
  RUN%FLCK         =  T
/

! -------------------------------------------------------------------- !
! Define the timesteps parameterization via TIMESTEPS_NML namelist
! -------------------------------------------------------------------- !
&TIMESTEPS_NML
  TIMESTEPS%DTMAX        =  120.
  TIMESTEPS%DTXY         =  40.
  TIMESTEPS%DTKTH        =  60.
  TIMESTEPS%DTMIN        =  10.
/

! -------------------------------------------------------------------- !
! Define the grid to preprocess via GRID_NML namelist
! -------------------------------------------------------------------- !
&GRID_NML
  GRID%NAME         =  'RRFS'
  GRID%NML          =  'namelists_curv.nml'
  GRID%TYPE         =  'CURV'
  GRID%COORD        =  'SPHE'
  GRID%CLOS         =  'NONE'
  GRID%ZLIM         =  0.0
  GRID%DMIN         =  1.0
/

! -------------------------------------------------------------------- !
! Define the rectilinear grid type via CURV_NML namelist
! -------------------------------------------------------------------- !
&CURV_NML
  CURV%NX              =  1799
  CURV%NY              =  1059
  CURV%XCOORD%SF        =  1.0
  CURV%XCOORD%OFF       =  0.0
  CURV%XCOORD%FILENAME  =  'lon.1799x1059.RRFS.dat'
  CURV%YCOORD%SF        =  1.0
  CURV%YCOORD%OFF       =  0.0
  CURV%YCOORD%FILENAME  =  'lat.1799x1059.RRFS.dat'
/

! -------------------------------------------------------------------- !
! Define the depth to preprocess via DEPTH_NML namelist
! -------------------------------------------------------------------- !
&DEPTH_NML
  DEPTH%SF        =  -1.
  DEPTH%FILENAME  =  'rrfs3km.depth_ascii'
/

! -------------------------------------------------------------------- !
! Define the input boundary points via INBND_COUNT_NML and
!                                      INBND_POINT_NML namelist
! -------------------------------------------------------------------- !
&INBND_COUNT_NML
  INBND_COUNT%N_POINT    =  6
/

&INBND_POINT_NML
  INBND_POINT(1)         =  102 86 F
  INBND_POINT(2)         =  102 118 T
  INBND_POINT(3)         =  2 118 T
  INBND_POINT(4)         =  2 2 T
  INBND_POINT(5)         =  102 2 T
  INBND_POINT(6)         =  102 11 T
/

! -------------------------------------------------------------------- !
! WAVEWATCH III - end of namelist                                      !
! -------------------------------------------------------------------- !

[JessicaMeixner-NOAA]

I don't see any red flags with this.